Automatic compensation for the light attenuation due to epidermal melanin in skin images
Melanin is an effective absorber of light and is able to dissipate the majority of light absorbed by the pigment. This can mask the absorbance of biomolecules of interest such as hemoglobin when imaging skin tissue. Melanin is the primary determinant of skin color which can vary between individuals as well as within individuals (due to the presence of portions of skin with more or less pigment as well as features such as freckles and moles). Described herein are methods for compensating for melanin absorption at the pixel level of an image, thereby allowing for more accurate imaging of concentration and distribution of biomolecules of interest in a tissue portion.
1. A method for compensating for melanin absorption during near-infrared imaging, visible imaging or near-infrared and visible imaging of concentration and distribution of biomolecules of interest in a tissue portion comprising:
determining skin color of the tissue portion, wherein the skin color is determined by measuring light reflected from the tissue portion at two or more wavelengths and using reflectance to calculate the skin color of the tissue portion;
determining a melanin correction factor from the skin color, wherein the correction for melanin absorption comprises multiplying absorption at each wavelength by the melanin correction factor and subtracting said absorption at each wavelength so as to subtract absorption due to melanin from the image; and
correcting for melanin absorption at each wavelength used for imaging the tissue portion, thereby providing a corrected image of the tissue portion and the biomolecules of interest within the tissue portion, and
displaying the corrected image of the tissue portion.
2. The method according to claim 1 wherein the subtraction is made at pixels of the image.
3. The method according to claim 1 wherein the skin color is calculated by a regression model.
4. The method according to claim 3 wherein the regression model is derived from a database of reflectance measurements from subjects with known skin color or skin melanin content.
5. The method according to claim 3 wherein the regression model is derived from the difference in the reflected at two or more wavelengths from a tissue portion to one or more prototype tissue portions with limited or extensive melanin.
6. The method according to claim 1 wherein the skin color is estimated from a color digital image of the tissue portion.